My 2 cents in the kitty.
Well this is my first post on this web site but, I've been around for a while. I also have built several ignitions, mind you they were for drag racing and msd inductive with 40 amp coilcharging. Anyway, I agree with John that a one-shot should be used in this circuit (timing) along with a lot more filtering around the cpu and on the HV board. I see some people are trying to make this work without using an oscilliscope which is almost impossible. I also agree with John that there is a bug in the code with regrards to the ringing when the magnet is parked under the hall effect.
The big thing about hall-effects is that they are not like a inductive pickup where there is a rise and fall in the signal and there is no signal if the magnet is not moving. The hall-effect will remain on so long as the magnet is at the pickup, so the cpu sees on,now is the cpu looking for a 'change' or just 'on'. If it is looking for 'on' then this could cause ringing. With MegaSquirt we get a lot of cpu resets from noise, I;m not saying this is a problem here but, from the schematics I've looked at there is almost no signal filtering on the HV board other than pull-up and pull-down resistors which will amplify any noise though the coupling cap. I would put a 5.6v zener between the signal and groundnext tothe scr gate to pull away over voltsand C7 should be a polarized cap tohelp with surges.
All ignition coils produce ringing and it is not always easy to get rid of. Some circuits make use of a 'clamp' or a low ohmage current resistor, what works best depends on the circuit. All hall-effects have a gauss (g)rating that determines when the switch will pickup and drop out, this in turn determines the strength of the magnet needed and the distance between sensor and magnet.
Part# /Operate Point (G) /Release Point (G)<table class="producttable"> <tbody> <tr class="alt"> <td>
A1120</td> <td><50 (typ 35)</td> <td>>5 (typ 25)</td> </tr> </tbody></table>
I hope I helped if not then it only cost me 2 cents.
Ray